Although mutations in cell populations and allelic frequencies in organismal populations are obviously not equivalent, the idea to halt therapy when some sensitive cells remain is really a fairly old idea: 40 years ago, with integrated pest management (IPM) came the slogan “leave a pest residue, not a pesticide residue”. And 20 years ago, resistance management was mandated for GMO Bt-crops; it called for “obligatory refuges”, where non-Bt-crops were planted to save the susceptible
alleles and thereby prevent fixation of homozygous resistant alleles. This practice has been incredibly successful––despite the fact that everyone predicted rapid resistance caused by continuous selection with the Bt-crop.
The opposite example is that of herbicide-resistant crops (“Roundup-ready” GMOs), where no resistance management strategy was used, resulting in increasing numbers of glyphosate-resistant weeds in the U.S. It simply seems unfortunate that the people studying resistance to anti-virals, antibiotics, cancer therapy, insecticides and herbicides, … so rarely read each others’ papers. The underlying biology may be somewhat different, … but “evolution by selection”, selective pressures, is always the central player.